EMO-R3：面向多模态大模型情感推理的反思式强化学习
EMO-R3: Reflective Reinforcement Learning for Emotional Reasoning in Multimodal Large Language Models
把可验证的反思式奖励（视觉-文本一致性 + 情感连贯性）接进 GRPO 类强化学习流程，为视觉-语言模型在主观、难以标注的任务上提供了一条可复用的训练范式；结构化思维链模板也可迁移到其他需要可解释推理的视觉理解任务。
原文依据 design a Reflective Emotional Reward that enables the model to re-evaluate its reasoning based on visual-text consistency and emotional coherence
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多模态大模型（MLLMs）在视觉推理与理解任务上已取得显著进展，但仍难以捕捉人类情感的复杂性与主观性。现有基于监督微调的方法往往泛化能力有限、可解释性较差，而 Group Relative Policy Optimization 等强化学习方法又无法与情感认知的内在特性相契合。为应对这些挑战，我们提出面向情感推理的反思式强化学习（EMO-R3），这是一个旨在增强 MLLMs 情感推理能力的框架。具体而言，我们引入结构化情感思维（Structured Emotional Thinking），引导模型以结构化、可解释的方式逐步进行情感推理；并设计反思式情感奖励（Reflective Emotional Reward），使模型能够基于视觉-文本一致性与情感连贯性重新评估其推理过程。大量实验表明，EMO-R3 显著提升了 MLLMs 的可解释性与情感智能，在多个视觉情感理解基准上取得了更优性能。
英文原文摘要 1185 字符
Multimodal Large Language Models (MLLMs) have shown remarkable progress in visual reasoning and understanding tasks but still struggle to capture the complexity and subjectivity of human emotions. Existing approaches based on supervised fine-tuning often suffer from limited generalization and poor interpretability, while reinforcement learning methods such as Group Relative Policy Optimization fail to align with the intrinsic characteristics of emotional cognition.To address these challenges, we propose Reflective Reinforcement Learning for Emotional Reasoning (EMO-R3), a framework designed to enhance the emotional reasoning ability of MLLMs. Specifically, we introduce Structured Emotional Thinking to guide the model to perform step-by-step emotional reasoning in a structured and interpretable manner, and design a Reflective Emotional Reward that enables the model to re-evaluate its reasoning based on visual-text consistency and emotional coherence. Extensive experiments demonstrate that EMO-R3 significantly improves both the interpretability and emotional intelligence of MLLMs, achieving superior performance across multiple visual emotional understanding benchmarks.